*Official* NBR Desktop Overclocker's Lounge [laptop owners welcome, too]

Well that "bug" exposes what is actually happening under the hood.

 

Thought it was a really cool response and a sign of good sportsmanship, but looks like he just deleted his reply:

https://twitter.com/aw_umar/status/1121081705547300864?s=19

"Oh damn"

 

Maybe try from Azor's account https://twitter.com/AzorFrank But yeah he removed his shocked statement.

 

i.e. it's not doing what the user is telling it to do!? Some kind of fail safes & 'protections'.

 

Seeing that I'm on the East coast, not likely. John is in their backyard but you never know. However, my local UPS service center despises me, due to me calling them out on their infamous non-delivery exceptions of no one being home when there are people waiting at that home. I wouldn't be surprised if I don't get my cards until the following week, lol.

 

Ugh I hate when that happens. I have everything delivered to my store open 9am-8pm. Always FedEx somehow manages to do the whole delivery exception when someone is always there. I normally get packages in my area from 11am-3pm.

 

It's reporting "fake values" to the OS.

 

https://hwbot.org/submission/4143596_

 

Hey guys, just a couple of updates. My son is home and doing well. He can even walk without a walker. Doc says no more hanging off high risers for a while though. He's a union glazier. Also my HWBOT was reinstated. Back on the team [ATTACH

 

@JoeT44 that's great news! Thanks for the update. Very glad to hear... nice scores too...

There is definitely something wonky going on for sure... frustrating lol... I've tested my 8700K, 8086K, 9700K and several 9900K's .... I agree in that it's something goofy at the firmware level... @ssj92 have you tried other CPU's?

I thought my KP's were arriving Tuesday, but it's actually tomorrow LOL... total silicon buzz kill .... oh well... I've been looking for my SLI Bridge hmm... it's in a box somewhere. Right when I reorder one I'll end up finding it...that's how it usually works right?

 

I don't have any other CPUs besides 3920XM, but I don't think this laptop can handle that CPU so it'll stay in my M18x.

 

If you want send me all firmware revisions and I'll take a look.

Also I wouldn't feel too bad about it.
Just checked and apparently this is the highest scoring Clevo without a Prema Mod and it's even using one of those "unlocked BIOS" as they like to call it these days:

https://www.3dmark.com/fs/19157795

 

Oh haha yea for sure... what has been your highest physics so far with the 9900K?

That'd be great. Thanks.

 

Are you referring to these?

1.5.0: https://downloads.dell.com/FOLDER05583129M/1/Alienware_Area-51m_1.5.0.EXE

1.3.2: https://downloads.dell.com/FOLDER05529946M/1/Alienware_Area-51m_1.3.2.EXE

1.3.1: https://downloads.dell.com/FOLDER05515804M/1/Alienware_Area-51m_1.3.1.EXE

1.3.0: https://downloads.dell.com/FOLDER05489398M/1/Alienware_Area-51m_1.3.0.EXE

1.2.0: https://downloads.dell.com/FOLDER05439585M/1/Alienware_Area-51m_1.2.0.EXE

1.0.1: https://downloads.dell.com/FOLDER05388803M/1/Alienware_Area-51m_1.0.1.EXE

Honestly, only 1.0.1, 1.2.0, 1.3.0 and 1.3.2 are worth looking at. Idk about 1.5.0 I haven;t flashed it yet.

1.0.1 Base release
1.2.0 added HT and a few OC options
1.3.0 increased power limit
1.3.2 was the bandaid fix for the blowing up 2080

I haven't done anything recently and am hoping to do so soon but I think these were my best?

Physics @ "5.1Ghz": https://www.3dmark.com/fs/18764713

GPU 28.7k: https://www.3dmark.com/fs/19051387

 

I think Prema needs packed archives with all BIOS EXE to check.

 

Ah kk..I had sent him a link to 1.3.2 earlier...

 

I mean all, from 1.00 to till date.. He will compare most bin files for changes using AMI BIOS developer tools for enhanced precision and debugging. He might not make a BIOS for AW then again any valuable hints will be good enough to break the limiting force.

 

All BIOS posted above use the same outdated microcode as the first IOI version.
For 9900K update the 906EC microcode from A2 to AE manually via Windows.
Also use Inspectre to disable Spectre and Meltdown protection and see what that does to your Physics score.

 

Ahhh very interesting... It had crossed my mind before the newer bios was released, but because it indicated that the micro-codes were updated, it didn't cross my mind to question that... hhhmmm very interesting.... Thanks for the heads up!

 

Latest BIOS only added support for KFC, they didn't update code for current chip models.

 

Do you still support skylake based Clevos? I did find that latest Microcode CC for 506e3 is bricking my AW throttlebook and wiped out CMOS with the infamous 7 beeps with no display. Had to backtrack to C2.

 

Usual Dell MO since 8 years....

 

SKL CC is working fine here on Clevo.

 

Hmmm... Same uCode as platomav's? https://github.com/platomav/CPUMicr...lat36_ver000000CC_2019-04-01_PRD_DFA96B26.bin
I used MMTool 5.0.0.7 it said saving as unsigned BIOS and I tried MMtool5.02 which saved it as secure BIOS and that's the one which caused bricked mobo. Any idea?

 

Hmm... makes sense... I hope I wasn't thinking of something else lol... I thought the latest indicated that the micro-code was updated... now I'm really curious lol ...

 

Yes, first of April CC revision. Maybe it's Intel fooling us.

No version of those tools will secure sign a BIOS. One version is just telling you that making changes will invalidate the signature while the other one doesn't.

 

Hmmm.... The old version said it is saving new BIOS as unsigned after uCode patching While new version did say saving as signed BIOS. So, I take it 506e3 is CC and 506e2 is E2 to be imported to BIOS, correct? Let me check if BIOS 1.6.1 AW have new certificate issued by Dell that didn't get expired.

 

Good morning fellas...

Physics Tune in Progress...: 28135
(Finally getting there... broke the 28K barrier...)



As much as I'd like to dive into benching the A51M...the KP's arrive today so that may take priority.

 

@iunlock
Now comes the balancing act.

 

Yes indeed. Precisely.

How's the KP doing on your end? Having lots of fun? I think I'll run them solo first ... had to order a SLI bridge anyway so it's not here yet.

 

@iunlock That is one cherry 9900K you have there. I might pick up a 9900KFC since I heard they overclock better.

My new personal best, was also just a first run so now that I'm home from vacation I can get to really testing this card. I think I need to remove the power limit of this FTW3 vBIOS. I think maybe the XOC vbios might help since this is without the 1.093v v/lock and just a straight +150Mhz core. Maybe a full block and rad for the card to keep the temps in check as well.

https://www.3dmark.com/3dm/35823755?

 

Very nice and welcome back. Hope you had fun on your vaca. I'll be leaving for a mini vaca in about a week or so ... trying to get as much benching and system stuff done before I head out...

Yea the 9900KFC looks pretty promising... pretty neato, especially with it not having the useless igpu haha ...

This cherry 9900K is still not delidded yet... I've been a bit delid shy to mess with this one haha... Although I've delided more cpu's than I can remember, it's just one of those things where the only time I hold my breath is when delidding cherry chips...

Looking forward to your benching! Cheers.

 

I haven't had a lot of time to really play with it, but have been trying to see if it will show higher than 1.093V ( For Brother @Mr. Fox ) in windows or on the digital display. It shows the voltage is going up using the DMM and Precision X1, but not in GPUZ, Hwinfo or MSI Afterburner. Voltage can run in the 1.5V range, but you will fry the card on water. Watt's as high as 440W, but the gains seem minimal. My card will do 2220 to 2035 core and 8575 - 8600 memory, but getting it to run optimal is the challenge it would seem. Max watts is 520 so plenty of room there, but I suspect it's too warm for the card while sitting on the AIO water setup. it needs to be in the -40 range to push into the 2300 Mhz range without losing performance.(Speculation of course)

 

Thanks for checking, Brother John. Interesting. I run 2220 on core all the time with stock voltage. If you're not able to run higher than 2220, maybe it's good that I didn't buy a K|INGP|N card. That would have been extremely upsetting to spend $1900 and not be able to overclock any higher.

So, bumping the voltage up to like 1.200V won't let you run 2300 on core?

I guess my biggest issue is my Micron GDDR6 limits my memory overclock by quite a bit.

NVIDIA is totally ruining the sport of overclocking with their stupid control freak bullcrap.

 

Very cool. I have a buddy working on 2310 on DICE... pretty impressive to see how strong these cards are. That's odd how it's not showing in some apps and not in others.. hmm... perhaps conflicting sensor data affecting the read outs?

I'll see how far I can push it with the stock AIO... it'll be fun. I wonder when the water block will be releasing?

Oh and with your core did you mean 2220 to 2335? - Wow that memory at 8600 ... Samsung FTW. What a monster.

Will you be going LN2 at all?

 

Some of the newer 2080Ti FTW3's have Samsung chips... my current card has micron (eww) so I'm in the same boat... I just wish they would rid micron all together and just stick with Samsung chips... at least with the higher end brands like EVGA. They can still bin the Sammy chips like they are already, that's fine, but just rid micron off the entire line of cards in the future....that'd be great.

 

Yes, and one of the issues is you don't know until you get the GPU. I doubt returning it or exchanging it for one with Samsung memory would be viewed as acceptable by the vendor. It's pretty discouraging at this point. Still thinking about selling my GPU and using the money to buy 2 or 3 cheap AMD GPUs. That, or just stop benching and stop using Windows. I'm ready to move 100% to Linux but for benching. If I were no longer benching there would be no reason for me to continue using Windows.

I have no issues with Micron per se. My GPU functions well in all respects. I'm not a Samsung (or any other brand) fanboy. My issue is only with respect to the memory overclock being limited. It only so happens that Samsung's name is on the chip that overclocks higher. I don't care whose name is on the parts as long as I get what I want.

 

Vega 56 is supposed to be fun to tweak. I mean you're not gonna score highly with it in comparison to a 2080ti, but it's more tweakable. Radeon VII, I don't remember that being very overclock friendly from anything I've seen.

 

Yes, I am aware... sadly. And, pretty much anything from AMD would be a downgrade in terms of results. But, I could still rack up hardware points even with sucky scores compared to the 2080 Ti. Radeon VII overclocks better than any GPU from AMD I have seen for years. Their GPUs haven't been good at overclocking for a very long time. I've been looking at results at 3DMark.com and there is nothing they bring to the table that can come even a little bit close to achieving 2080 Ti performance. Doing this would be mostly for the personal satisfaction of giving NVIDIA the nasty-finger salute for their GPU castrating nonsense. I have very little respect left for them. They suck. It's unfortunate they still hold the performance crown, even with their artificially crippled products running cancer firmware.

 

Yea that is what makes it no fun in not knowing what you're going to get. The only sure way is to buy it from a trusted third party seller... but anyhow, the reason I mention Samsung is because there are no other companies making GDDR6 for our cards, but micron and Samsung at the moment. That may change soon in the near future as SKHynix is going to be joining the party. It's not about brand fanboyism; there was no mention of that whatsoever...what it's about is dealing with the only two who makes the chips..micron and Samsung...

Window's wise, it's nothing but a platform for gaming and benching for me as well... the daily driver has always been Linux ... I can't wait to see what comes out of the gpu drivers when the code goes open source here soon...

 

No problem.


Is about where I am now using AIO.
Voltage on card shows what PX1 shows while a real dmm shows something different.

Unfortunately no one is running an optimal 2300 anything on water. Unless it's a modified chiller running below -30C

And Nvidia has pretty much stayed true to form with control, that's why all the extreme measures come into play.

Yes, i'm hoping for more on a standard aio setup, but we shall see if there is any real gain as time progresses.

Not sure about the software, but this card has true read points built into the card. which tell the truth. PX1 will also tell the the truth mostly. anything else with a KPE will tell stories.

From the rumour mill, (jacob) they said they are not making a hydrocopper version, but there will be a water block....

Edit: 2200 - [email protected] (now seems stable) on core. And also utilization seems to be low @ 94%

 

That's excellent information. Thanks for sharing. I'm not nearly as disappointed that my wife pitched a fit and said no now. Based on data in that table I am pretty sure I would have regretted buying one.

Not saying that you're a fanboy. The only reason I mentioned the fanboy thing was because of the "eww" comment about Micron. I have no issues with Micron as a brand and just wanted to make that aspect of it clear. If I were a gamer more than an overclocker it would not matter one iota and I would not care. The only reason it matters is the overclocking capabilities of the Samsung GDDR6 is higher. That could change at some point.

Edit: Now, if I were inclined to make choices based on brand and principle rather than results, I would definitely choose Micron on that basis since they are an American company and HQ'd in Boise, ID.

 

Yes, I believe so. Unless one is ready to go extreme, it performs somewhere along the rest....
2250 and 2265
https://www.3dmark.com/compare/fs/19222260/fs/19222285

 

And the problem I have with doing it this way is this....
https://www.3dmark.com/3dm/35826050

 

Yes, that graphics score is not as high as what it should be for the overclock. Seems way low. I ran this last week... https://www.3dmark.com/fs/19158197

 

It's way to low vs this one i just ran another way.
https://www.3dmark.com/3dm/35826357

And now this one.
https://www.3dmark.com/3dm/35826479

Interesting that it shows triple channel...

 

What did you do differently, other than lower the core clock?

Yeah, weird it shows triple channel. I wonder what the deal is with that? Are the memory speeds the same as quad channel, or did the read/write/copy speeds go down as well?

 

This one went up in gpu and dropped in cpu and combined
Didn't use the voltage offset.

 

That looks like fun!

 

...and disappointment at the same time at least for non-exotic cooling haha

Sent from my Xiaomi Mi Max 2 (Oxygen) using Tapatalk